Update Pair Extraordinaire Counts #247
Conversation
After having set up a script to test for myself, I have found that the actual counts to get the pair extraordinaire levels are as edited above.
|
Very interesting! Because of #137: @julthep seems like the count is different after all! CC @Rongronggg9 I'll leave the PR open for a bit for all to react, but automated testing is of course great proof for the numbers! |
Yes I plan to leave https://github.com/nathanielop/achievements available, as well as update it when possible. I'm still going through it to make it a bit more robust and customizable for people, however. Figured there were probably other semi-completionists out there (due to the 2020 Mars and Arctic Vault achievements being unobtainable) who would find a tool like that handy. |
|
Amazing! I have a doubt if the count is summed according to the count of coauthored users per PR instead of the count of coauthored PRs, considering this is the only leveled achievement that comes without statistics. That may explain why the time between my |
Could very well be. Each of my PRs were coauthored by two other people and I counted commits only (seems based on commits vs PRs? otherwise my numbers would be completely random seemingly). For the most part each of my coauthored commits were individual commits + PR merges pairs versus several commits under a single PR merge. At most, changing from commits to PRs, my total count might go down to between 90 and 100? The levels of 1 -> 15 -> 30 -> 120 coauthored commits feels like a lot better of a fit. https://github.com/DIYgod/RSSHub/pull/9761/commits of yours seems to have ~40 coauthored commits on it alone? As stated though, there's nothing really provided on githubs part so it's kind of guess and check, just hoping my somewhat more controlled numbers (albeit not controlled enough to provide any closure for certain) helps narrow it down. |
Probably! The achievement history list all users who have ever committed to the PR, which makes me misunderstand the condition of obtaining it. But let us reexamine its description " |
|
Maybe you could introduce a parameter to the Pair Extraordinaire obtainer script to control the number of coauthored users and recruit some volunteers who have not got Pair Extraordinaire for research? |
|
Does the more explicit wording mean one could do 120 commits in one PR? 🤔 |
I believe so, yes. I'll need to do what @Rongronggg9 mentioned to know for sure but if I just go by PRs then my numbers end up more like 1 -> 7 -> 24 -> 93 which doesn't seem very likely to be the basis for leveling the badge. |
Yes, that's actually what https://github.com/nathanielop/achievements/blob/a31afebb3109ddaa4ecf6196fed8a51e5b266829/bin/pair-extraordinaire does, even from its initial version. But we just need a volunteer to examine it... |
|
I'll test this later tonight, I have a volunteer. I'll have to give setup some thought but I'd like to test commits versus PRs and coauthor count effects, although I'm unsure if I'll be able to independently test all 3 of these variables in the small span between level 0 and level 4 of pair extraordinaire |
|
Hi, Im also interested in your code and improve it with some fixes and tests. about P-E, I tried to make a commit with @nathanielop per PRs and it seems that it counts the number of PRs. I accidentally set the sleep to 20, so my code only ran about 30-40 times before rejected :) |
|
@MasterPi-2124 Would you please slightly modify the script to add one more coauthored person (not me, plz), to see if the number of coauthored person(s) matters? 🤔 |
Yeah, I will make a bash that asks if you want to add your friend to the commits and check, but you should ask them before :)) Regarding other achievements, I and @nathanielop will take a look ad see if we can try to earn them. |
|
@Rongronggg9 This marked my P-E x3. |
|
Wait, I made a mistake. MasterPi-2124/achievements#42 with 21 coauthored commits only made @MasterPi-2124 earn P-E x1, and the condition of earning x1 should be 1 coauthored commit (instead of 15, I misread). And after merging 9 additional PRs with 1 coauthored commit each, @MasterPi-2124 earned P-E x2 with 30 coauthored commits. But the condition of earning x2 in this PR is 15 (instead of 30, I misread), which means that @MasterPi-2124 should have earned their P-E x2 with MasterPi-2124/achievements#42. That contradicts @nathanielop's conclusion... |
@MasterPi-2124 I count your PRs. MasterPi-2124/achievements#42 contains 21 coauthored-with-1-person commits. 19 PRs contain 1 coauthored-with-1-person commit each. 5 PRs ( |
|
I think we should abandon the concept I originally presented with this PR which was that the achievement is based on commits instead of merged PRs. While I haven't had much chance to test due to the painful process of changing my local git setup to point to a testing account, I did test my 120 coauthored commits in a single PR theory, wherein I found out I was wrong. That being said, PE 1x still only requires a single coauthored commit/pr merged, as proven via this PR on my achievements repo, where you can see a single commit/merged PR give my tester account PE 1x. I'm in the process of modifying me and @MasterPi-2124 achievements scripts to granularly increment PR counts and ensure a single commit per PR to allow for a more controlled method of verifying our numbers. After that's done, I can reproduce on a separate account with more than one coauthor to ensure coauthor counts don't change anything. Have to say I'm rather surprised by the complexity behind this seemingly straightforward achievement. |
|
@nathanielop I noticed that @achiev3 and you share the same email domain. As a reminder, if that account is yours, this may violate the ToS of GitHub:
https://docs.github.com/en/site-policy/github-terms/github-terms-of-service#3-account-requirements |
Yeah, I was aware of this, but for lack of an alternative figured I'd use it as a test bed, and delete it afterwards. I assume the more general reason for that provision in the ToS is to prevent spammers and star bots or otherwise. I do not intend to keep the account beyond the purposes of testing here. Besides, it's really closer here in usage to a machine user than an actual user as all actions of it are essentially automated through the scripts in the achievements repo. |
|
We, ani-cli, might have another interesting datapoint for you shortly to verify the thesis that the achievement is counted by coauthered commits rather than PRs. |
Does not seem to work. Just rebase-merged a PR with 40 commits, all coauthored by 5-6 people. Only got pair extraordinaire x2 Edit: my co-maintainer that had x2 before now has x3 |
|
Just did a follow-up PR with one coauthored commit (same 5 people) 71zenith got bumped again to x2, but the rest of us didn't get bumped. Weird counts... |
|
Maybe 16-64-128? |
|
I am not sure how it works but I did get the bronze and silver at 10 and 24 PRs respectively so I guess I could confirm those counts. However, I am now at 52 and I haven't received the gold one, so I don't really know what to expect. |
|
I'll test this later tonight, I have a volunteer. I'll have to give setup some thought but I'd like to test commits versus PRs and coauthor count effects, although I'm unsure if I'll be able to independently test all 3 of these variables in the small span between level 0 and level 4 of pair extraordinaire |
|
After all this time, this remains a mystery. Will I ever reach pair extraordinaire x4, aka heaven? |
|
After having set up a script to test for myself, I have found that the actual counts to get the pair extraordinaire levels are as edited above.